Introduction Beneficial effects of physical activity (PA) on health are well known, but limited evidence exists for lifecourse factors that may influence physical activity levels in adulthood. Our aim was to investigate the relative contributions of a range of factors from across life to variations in PA levels in the Newcastle 1000 Families Study.

Methods Detailed information was collected prospectively during childhood. At age 50 years, 574 study members returned self-completion questionnaires. Responses included details of PA levels, collated into four indicators—work, commuting (walking, cycling), household (house work, DIY, gardening) and sport activities. Each was defined by three categories (inactive, less active, most active) and analysed by ordered logistic regression, factor analysis and path analysis.

Results Males had higher levels of work activity (p=0.010) and lower levels of household activity (p<0.001). Increased sport (p=0.009) and household (p=0.002) activity were associated with reduced BMI. Increased sport activity was associated with more advantaged social class (p=0.004) and an increase in work activity with middle classes (p<0.001). Current smoking was associated with reduced sport (p<0.001) and work (p=0.012) activities and higher achieved education with decreased household and commuting activities. Factor analysis identified two components: “sport” and “all other PA”, to be considered separately in path analysis which showed highest relative contributors were BMI for other PA and current social class and smoking for sport.

Conclusion The association between PA and BMI should be considered bidirectional. Early life factors were negligible when contemporary factors were addressed with the exception of achieved education.